Revisiting Banker Pay
Via Tyler Cowen, I noticed this morning that Jeffrey Friedman had written a blog post suggesting that executive compensation practices weren't really at the core of our recent financial crisis. Since I'm sort of a squish on this subject myself, I clicked over to read it. Unfortunately, this was a big part of Friedman's case:
Perhaps the most powerful evidence against the executive-compensation thesis, however, is that 81 percent of the mortgage-backed tranches purchased by banks were rated AAA, and thus produced lower returns than the double-A and lower-rated tranches of the same mortgage-backed securities that were available. Bankers who were indifferent to risk because they were seeking higher return, hence higher bonuses, should have bought the lower-rated tranches universally, but they did so only 19 percent of the time. And most of those purchases were of double-A rather than A, BBB, or lower-rated, more-lucrative tranches.
Oh come on. Before 2001, banks were required to meet full capital requirements for all asset-back securities on their books. But in January 2002 that changed: AAA-rated securities, because they were so safe, were allowed to be backed by only 20% of the usual capital. That meant banks had every incentive to manufacture and keep on their books as much AAA debt as possible. It allowed them to increase their leverage fivefold.
Later, of course, regulators went further and allowed banks to use their own internal models for risk adjustment, which gave them even more incentive to play games with risk weighting. As a result, they sliced and diced their securities into a little bit of lower-rated stuff and a whole lot of super-senior tranches, which their models said were so safe they required barely any capital at all to back them. Leverage went through the roof.
Banks didn't hold lots of AAA-rated securities because bankers were playing it safe. They held lots of AAA because it allowed them to game their capital requirements and pile ever more risk on their books. It's evidence of exactly the opposite of what Friedman suggests.
POSTSCRIPT: That said, I continue to think that leverage and capital requirements are the key issues we should be concerned about, not compensation per se. Bear and Lehman didn't collapse because they made bad bets on the housing market. They collapsed because they made bad bets on the housing market backed by truckloads of overnight repo financing. When the financing dried up, they went under. If they hadn't overleveraged their bad bets, they'd be weaker today but still around. Likewise, the financial system would be licking its wounds but not responsible for a global meltdown.